Adds basic single-node support. This is rebased off of #260 so it's only the last ~13 commits that are relevant. This will probably get broken down into 2 or 3 PRs because there were some reorganization of the network and storage questions.
This adds an initial question to microcloud init asking if the user wants to set up just one node.
Do you want to concurrently set up more than one cluster member? (yes/no) [default=yes]:
If the user enters no, the system will follow the ordinary setup with ZFS and OVN (or optionally a FAN network instead), but the user will not be prompted for a lookup subnet and will not be presented a list of systems.
MicroCeph will be skipped as we require 3 systems. MicroCeph does not yet support a way to modify OSD configuration from the API, so we cannot change replication rules just yet.
The cluster can be grown after this point by using microcloud add.
A few changes were made to the disk and network selection to inspect LXD's storage pools and networks to determine if the existing local node should be included in the list for disk selection.
TODO:
[ ] Add tests
[ ] Support preseed
[ ] Support adding missing services (maybe, if it makes sense)
Adds basic single-node support. This is rebased off of #260 so it's only the last ~13 commits that are relevant. This will probably get broken down into 2 or 3 PRs because there were some reorganization of the network and storage questions.
This adds an initial question to
microcloud initasking if the user wants to set up just one node.If the user enters
no, the system will follow the ordinary setup with ZFS and OVN (or optionally a FAN network instead), but the user will not be prompted for a lookup subnet and will not be presented a list of systems.MicroCeph will be skipped as we require 3 systems. MicroCeph does not yet support a way to modify OSD configuration from the API, so we cannot change replication rules just yet.
The cluster can be grown after this point by using
microcloud add.A few changes were made to the disk and network selection to inspect LXD's storage pools and networks to determine if the existing local node should be included in the list for disk selection.
TODO: